How this instrument works
A z-score tells you how many standard deviations a value sits from the mean; the raw score calculation reverses that. Given a target z (say, you want to know the value that sits exactly one standard deviation above average), and the mean and standard deviation of the group in question, this calculation returns the actual raw value on the original scale — test points, millimetres, dollars, whatever the underlying measurement is.
This reverse direction comes up constantly in practical settings that have nothing to do with checking an existing data point. A teacher setting grade-band cutoffs might decide the honors threshold should sit at z = +1.5 rather than picking a raw score out of the air; a quality-control engineer might define a process control limit at z = +3; a researcher designing a study might need to know what raw measurement corresponds to a chosen z-cutoff before collecting any data at all. In each case, the z-score is decided first, as a policy or a statistical choice, and the raw score is what you compute afterward to make it concrete and usable.
The underlying relationship is exactly the same one the z-score formula describes, just solved for a different variable — algebraically rearranging z = (x − μ) / σ to isolate x gives x = μ + z×σ. Nothing new is being claimed statistically; it's the identical relationship between a value, its mean, and its spread, read in the opposite direction.
- Enter your target into Z-score — how many standard deviations above (positive) or below (negative) the mean you want the raw value for.
- Enter the group's average into Mean (μ).
- Enter the group's spread into Standard deviation (σ) — it must be greater than zero.
- Read Raw score (x) — the actual value, on the original measurement scale, that corresponds to your chosen z-score.
- To go the other direction — you already have a raw value and want to know its z-score — use this site's z-score calculator instead.
Worked example — setting a cutoff one standard deviation above average
A teacher decides that the top grading band should start exactly one standard deviation above the class average, on a test where the mean is 75 and the standard deviation is 10 — a policy choice of z = 1.0, made before looking at any individual score. Enter 1.0 into Z-score, 75 into Mean (μ), and 10 into Standard deviation (σ).
The instrument computes x = 75 + 1.0 × 10 = 85. Any student scoring 85 or above on this test meets the z = 1.0 threshold the teacher set — the abstract 'one standard deviation above average' target has been converted into a concrete, postable cutoff score.
Questions
Why would I start from a z-score instead of a raw value?
Because sometimes the target is naturally expressed on the standard-deviation scale rather than the original one — a grading policy set at 'one SD above average,' a quality-control limit set at 'three sigma,' or a research design that specifies a cutoff in standardized terms before any data exists. This calculator converts that abstract target into a concrete number on the scale you actually work in.
What does a negative z-score produce here?
A raw score below the mean. Entering z = −1.5 with the same mean-75, SD-10 example gives x = 75 + (−1.5)×10 = 60 — a cutoff or reference point below average by one and a half standard deviations. The sign of z always determines whether the resulting raw score lands above or below the mean.
What happens if I enter z = 0?
You get back the mean exactly, every time — x = μ + 0×σ = μ. A z-score of zero always corresponds to the mean itself, regardless of what the standard deviation is, since the spread term drops out of the calculation entirely.
Is this just the z-score formula in reverse?
Yes, exactly — z = (x − μ)/σ rearranged to solve for x instead of z. They are true inverses: run this calculator's output back through the z-score calculator with the same mean and standard deviation, and you'll get your original z-score back. The two pages exist separately because the two questions ('how unusual is this value?' vs 'what value hits this target?') come up in different situations, even though the arithmetic is identical.
Does the standard deviation need to be from the whole population?
No — use whichever standard deviation actually describes the group your mean is drawn from, population σ or sample s. The formula treats it the same way either way; what matters is that the mean and standard deviation you enter both describe the same group consistently.
Can I use this for percentile-based cutoffs, like 'top 16%'?
Only after converting the percentile to a z-score first, since this calculator takes z as its input, not a percentile directly. For a roughly normal distribution, the empirical rule gives some standard landmarks — the top 16% corresponds to about z = 1.0, and the top 2.5% to about z = 2.0 — but for precise percentile-to-z conversions you'd want a standard normal table or a dedicated inverse-normal calculation.